
# Importing libraries
import pandas as pd
import numpy as np
from dataprep.eda import create_report
import datetime as dt
import warnings
warnings.filterwarnings('ignore')
warnings.warn('ignore')


# Using Pandas to read the dataset.
df = pd.read_csv("dublinbikes_20210401_20210701.csv")
# Using head() to visualize the top 5 rows
df.head(5)


# Data Type of each variable.
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2884576 entries, 0 to 2884575
Data columns (total 11 columns):
 #   Column                 Dtype  
---  ------                 -----  
 0   STATION ID             int64  
 1   TIME                   object 
 2   LAST UPDATED           object 
 3   NAME                   object 
 4   BIKE STANDS            int64  
 5   AVAILABLE BIKE STANDS  int64  
 6   AVAILABLE BIKES        int64  
 7   STATUS                 object 
 8   ADDRESS                object 
 9   LATITUDE               float64
 10  LONGITUDE              float64
dtypes: float64(2), int64(4), object(5)
memory usage: 242.1+ MB


# df.describe is a way to summarize the data using Descriptive Statistics: 
#Central Tendency and Measures of variability for Numerical features (except Station ID and Latitude and Longitue)
# I am Using .round() to suppress scientific notation.
df.describe(include=['int64']).round()


# Using df.describe to get summary statistics for objects
df.describe(include=np.object)


# Converting Last_Updated to DateTime type.
df['LAST UPDATED'] = pd.to_datetime(df['LAST UPDATED'])
df['TIME'] = pd.to_datetime(df['TIME'])


# Using EDA Dataprep to create a full report about the Dataset.
create_report(df).show();


df.groupby('BIKE STANDS').NAME.nunique()

BIKE STANDS
16     2
20    16
21     1
22     1
23     2
24     1
25     2
27     1
29     4
30    27
31     1
32     1
33     1
35     2
36     2
38     3
40    42
Name: NAME, dtype: int64


# The probability of exactly 3, when n=5 and p= 0.25 using python
from scipy.stats import binom
binom.pmf(k = 3, n = 5, p = 0.25 ).round(2)

0.09


#Importing library
import scipy
#To find the probability that the variable has a value GREATER than or equal to ussing SF Survival Function
scipy.stats.norm.sf(15,20,10).round(2)

0.69


#Creating a new dataset with unique values of each variable.
df1 = df[["NAME","LATITUDE","LONGITUDE"]]
df1 = pd.DataFrame(df1.drop_duplicates(subset = "NAME"))
df1.reset_index(drop = True, inplace = True)


#Plotting Station Map
# import folium
import folium

# create a base map centered around Dublin
mapObj = folium.Map(location=[df1.LATITUDE.mean(), df1.LONGITUDE.mean()], zoom_start=13.5, control_scale=True)


#creating lists to storage observations values separately
longs = df1["LONGITUDE"]
lats = df1["LATITUDE"]
names = df1["NAME"]

# create marker object for Dublin, one by one for every location in data DataFrame
for i in range(0,df1.shape[0]): # .shape[0] for Pandas DataFrame is the number of rows
    
    # create marker for location i 
    markerObj = folium.Marker(location = [lats[i],
                                          longs[i]],
                                          tooltip=names[i],
                                          icon=folium.Icon(color="blue",
                                          icon='bicycle', prefix='fa'))
    
    # add marker to map
    markerObj.add_to(mapObj)
# display map
mapObj


# Converting Last_Updated to DateTime type.
df['LAST_UPDATED'] = pd.to_datetime(df['LAST UPDATED'])


# Series.dt.dayofweek The day of the week with Monday=0, Sunday=6.
df['DAY_NUMBER'] = df.LAST_UPDATED.dt.dayofweek
# Adding a day type collumn, if weekday or weekend
df['DAY_TYPE'] = np.where(df['DAY_NUMBER'] <= 4, 'Weekday', (np.where(df['DAY_NUMBER'] == 5, 'Saturday', 'Sunday')))


# Getting a new column for Time rounded each 10 minutes
df["TIME_ROUNDED_10_MIN"] = pd.DataFrame(pd.Series(df["LAST_UPDATED"]).dt.round('10min'))
df['NEW_TIME'] = [dt.datetime.time(i) for i in df["TIME_ROUNDED_10_MIN"]]


#creating important feature Occupancy %
df['PERCENTAGE_OCCUPANCY'] =  df['AVAILABLE BIKES'] / df['BIKE STANDS']


# Removing redundant columns
df_clean = df.drop(['STATION ID','TIME','LAST_UPDATED','ADDRESS','LATITUDE','LONGITUDE','STATUS','LAST UPDATED'], axis=1, inplace=False)


df_clean.head(5)


# Using hierarchy to cluster the observations
#importing the library
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
#recovering matplotlib defalts
#plt.rcParams.update(plt.rcParamsDefault)

#Importing scipy.cluster.hierarchy to create dendrogram
import scipy.cluster.hierarchy as sch

dendrogram =sch.dendrogram(sch.linkage(
                df_clean.pivot_table(
                "PERCENTAGE_OCCUPANCY", "NEW_TIME", "DAY_NUMBER", 
                aggfunc="mean").rename_axis(columns=None).T
                ,method = 'ward'))

#creating the plot, assigning labels and title
plt.title('Dendrogram')
plt.xlabel('Week Days')
plt.ylabel('Euclidean distances')

# Add horizontal line to visualize .
plt.axhline(y=0.068, c='red', lw=1, linestyle='dashed')
plt.axhline(y=0.05, c='grey', lw=1, linestyle='dashed')

plt.show()


# Ploting a Time series in a Line Chart
# Importing library
import plotly_express as px

# Creating the plot
fig = px.line(df_clean.pivot_table(
                "PERCENTAGE_OCCUPANCY", "NEW_TIME", "DAY_TYPE", 
                aggfunc="mean").rename_axis(columns=None), 
              x=df_clean.pivot_table(
                "PERCENTAGE_OCCUPANCY", "NEW_TIME", "DAY_TYPE", 
                aggfunc="mean").rename_axis(columns=None).index, 
              y= df_clean.pivot_table(
                "PERCENTAGE_OCCUPANCY", "NEW_TIME", "DAY_TYPE", 
                aggfunc="mean").rename_axis(columns=None).columns,
             title="Dublin Bike Stations- Mean Occupancy per hour by day Type")
# Set custom plot
fig.update_layout(
    legend_title = "Day Type",
    title_font_color="gray",
    legend_title_font_color="black",
    xaxis_title= 'Time',
    yaxis_title= "Occupancy %")

# Set custom x-axis labels
fig.update_xaxes(
    ticktext=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00",
              "12:00:00","14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
    tickvals=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00","12:00:00",
              "14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
)
fig.update_xaxes(tickangle=90)

fig.show()


# Creating a new Data Frame with only Weekdays Monday to Fidray
new_df= df_clean[df_clean["DAY_TYPE"] == "Weekday"]
new_df.head(5)


# Importing library
import plotly_express as px
fig = px.line(new_df.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "NAME", 
                                 aggfunc="mean").rename_axis(columns=None), 
              x=new_df.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "NAME", 
                                 aggfunc="mean").rename_axis(columns=None).index, 
              y= new_df.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "NAME", 
                                 aggfunc="mean").rename_axis(columns=None).columns,
             title="Dublin Bike Stations- Workdays Mean Occupancy% by hour ")
fig.update_layout(
    legend_title = "Stations",
    title_font_color="gray",
    legend_title_font_color="black",
    xaxis_title= 'Hour',
    yaxis_title= "Occupancy %")
# Set custom x-axis labels
fig.update_xaxes(
    ticktext=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00",
              "12:00:00","14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
    tickvals=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00","12:00:00",
              "14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
)
fig.update_xaxes(tickangle=90)
fig.show()


#Creating a new Dataset
station_dataset = new_df.pivot_table("PERCENTAGE_OCCUPANCY","NAME", "NEW_TIME",aggfunc="mean").rename_axis(columns=None)
station_dataset.head(5)


#importing the libraries 
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler


# Use StandardScaler from Scikit Learn to standardize the features onto unit scale (mean = 0 and standard deviation = 1) 
# Thi is a requirement for the optimal performance of many Machine Learning algorithms including PCA.
data_scaled = StandardScaler().fit_transform(station_dataset)


# How many components are needed to explain 90% of the variance ?
pca = PCA(.90)
# fit and transform our dataset
fit = pca.fit_transform(data_scaled)
#print the number of components needed
pca.n_components_

2


# How much of the variance each component can explain ?
ratios = pca.explained_variance_ratio_
ratios

array([0.68027642, 0.23904475])


# keep the first two principal components of the data
pca = PCA(n_components=2)

# fit PCA model
principalComponents = pca.fit(data_scaled)

# transform data onto the first two principal components
data_pca = pca.transform(data_scaled)
#printing shape results
print("Original shape: {}".format(str(station_dataset.shape)))
print("Reduced shape: {}".format(str(data_pca.shape)))

Original shape: (109, 144)
Reduced shape: (109, 2)


df2 = pd.DataFrame(data_pca)
df2.columns = ['First_Comp','Second_Comp']


# Importing Library
import seaborn as sns
import matplotlib.pyplot as plt
#creating a pairplot to visualise our Clusters
_ = sns.pairplot(df2);
#add overall title
_.fig.suptitle('PCA 2 Principal Components');
#move overall title up
_.fig.subplots_adjust(top=.9);


# Using hierarchy to cluster the observations
# importing the library
import matplotlib.pyplot as plt
#recovering matplotlib defalts
#plt.rcParams.update(plt.rcParamsDefault)

#Importing scipy.cluster.hierarchy to create dendrogram
import scipy.cluster.hierarchy as sch
dendrogram = sch.dendrogram(sch.linkage(data_pca, method = 'ward'))
#creating the plot, assigning labels and title
plt.title('Dendrogram')
plt.xlabel('Stations')
plt.ylabel('Euclidean distances')
# Add horizontal line.

plt.axhline(y=56, c='grey', lw=1, linestyle='dashed')
plt.axhline(y=75, c='red', lw=1, linestyle='dashed')

plt.show()


# Including Cluster Labels
# importing the library 
from sklearn.cluster import AgglomerativeClustering

hierarchy_cluster = AgglomerativeClustering(n_clusters=2, affinity="euclidean", linkage="ward")
#create a pandas series to be used latter on
y1 = pd.Series(hierarchy_cluster.fit_predict(data_pca), name="Cluster Hierarquical")


#creating a new dataframe
df2 = pd.DataFrame(data_pca)
df2.columns = ['First_Comp','Second_Comp']

#assigning the Hierarchy cluster labels for each observation
df2 = pd.concat([df2, y1], axis = 1)
df2.head(5)


# Plotting Hierarchy Clusters
#importing the library
import matplotlib.pyplot as plt
#creating the plot
_ = sns.pairplot(df2, hue = 'Cluster Hierarquical',palette=['b', 'r'])
#move overall title up
_.fig.subplots_adjust(top=.9)
#add overall title
_.fig.suptitle('Hierarchy Cluster');


# Plotting Elbow method
import matplotlib.pyplot as plt
#recovering matplotlib defaults
#plt.rcParams.update(plt.rcParamsDefault)

#Using the elbow method to find the optimal number of clusters
from sklearn.cluster import KMeans
wcss = []
for i in range(1, 11):
    kmeans = KMeans(n_clusters = i, init = 'k-means++', random_state = 0)
    kmeans.fit(data_pca)
    wcss.append(kmeans.inertia_)
plt.plot(range(1, 11),wcss,marker='o')
plt.title('The Elbow Method')
plt.annotate('Possible number of Clusters', xy=(6, 2500), xytext=(5.4, 8000), arrowprops=dict(facecolor='green'))
plt.annotate("",xy=(4, 4000), xytext=(5.2, 7900), arrowprops=dict(facecolor='green'))
plt.annotate("",xy=(5, 3000), xytext=(5.3, 7900), arrowprops=dict(facecolor='green'))
plt.xlabel('Number of clusters')
plt.ylabel('WCSS')
plt.show()


#Importing and Fitting K-Means to dataset
from sklearn.cluster import KMeans
kmeans = KMeans(n_clusters = 5, init = 'k-means++', random_state = 0)
y2 = pd.Series(kmeans.fit_predict(data_pca), name="Cluster KMeans")


#importing the library
from sklearn.metrics import silhouette_score

# Calculate Silhouette Score hierarchy
hierarchy_score = silhouette_score(data_pca, hierarchy_cluster.labels_, metric='euclidean')
# Print the score
print('Silhouette Score Hierarchy for 2 clusters: %.3f' % hierarchy_score)


# Calculate Silhouette Score kmeans
score = silhouette_score(data_pca, kmeans.labels_, metric='euclidean')
#print('Silhouetter Score for x clusters : %.3f' % score)

# Print the score
print('Silhouette Score KMeans for 4 clusters : 0.395')
print('Silhouette Score KMeans for 5 clusters : 0.422')
print('Silhouette Score KMeans for 6 clusters : 0.380')

Silhouette Score Hierarchy for 2 clusters: 0.358
Silhouette Score KMeans for 4 clusters : 0.395
Silhouette Score KMeans for 5 clusters : 0.422
Silhouette Score KMeans for 6 clusters : 0.380


# Creating a new df using the two principal components
df3 = pd.DataFrame(data_pca)
df3.columns = ['First_Comp','Second_Comp']

# Using concat function to add KMeans Cluster Label
df3 = pd.concat([df3, y2], axis = 1)
df3.head(5)


# Plotting KMeans Clusters
#Importing the library
import matplotlib.pyplot as plt
#creating a pairplot to vizualise our Clusters
_ = sns.pairplot(df3, hue = 'Cluster KMeans',palette=['blue', 'red','green','orange','0']);
#add overall title
_.fig.suptitle('Kmeans Cluster');
#move overall title up
_.fig.subplots_adjust(top=.9);


#Creating a new dataFrame called stations with station names
stations = station_dataset.index.to_frame(index=False, name="NAME")
# Adding Kmeans Cluster prediction collumns to station dataset using concat
station_labels = pd.concat([stations,y2], axis = 1)
station_labels.head(5)


# Merging station_labels with df1 in order to add the clusters label and re-plot our map
df4= pd.merge(station_labels, df1)


# Creating a function to add a new column to carry-out the color of each Kmeans cluster
# For good practice I am going to sort the colours in a way to match with the pairplot 
conditionlist = [
    (df4['Cluster KMeans'] == 0) ,
    (df4['Cluster KMeans'] == 1) ,
    (df4['Cluster KMeans'] == 2) ,
    (df4['Cluster KMeans'] == 3) ,
    (df4['Cluster KMeans'] == 4)
]
choicelist = ['blue','red','green','orange','black']
              
df4['KMeans Color']  = np.select(conditionlist, choicelist, default='Not Specified')
df4.head(5)


#checking cluster size
df4['KMeans Color'].value_counts(ascending=False)

green     37
black     24
blue      19
orange    17
red       12
Name: KMeans Color, dtype: int64


#Re-plotting Station Map by KMean Cluster
# import folium
import folium

# create a base map centered around Dublin
mapObj = folium.Map(location=[df4.LATITUDE.mean(), df4.LONGITUDE.mean()], zoom_start=13.5, control_scale=True)

#creating lists to store each observation values separately
longs = df4["LONGITUDE"]
lats = df4["LATITUDE"]
names = df4["NAME"]
colors = df4["KMeans Color"]

# create marker object for Dublin, one by one for every location in data DataFrame df4
for i in range(0,df4.shape[0]): # .shape[0] for Pandas DataFrame is the number of rows

    # create marker for location i 
    markerObj = folium.Marker(location = 
                              [lats[i],
                              longs[i]],
                              tooltip=names[i],
                              icon=folium.Icon(color=colors[i],
                              icon='bicycle', prefix='fa'))
    
    # add marker to map
    markerObj.add_to(mapObj)
# display map
mapObj


## Adding labels to the new_df
df5= pd.merge(df4,new_df, on="NAME", how= "left")
df5.head(5)


#Plot Clusters by Mean Occupancy %
import plotly.express as px
fig = px.line(df5.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "KMeans Color", 
                              aggfunc="mean").rename_axis(columns=None), 
              x=df5.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "KMeans Color", 
                                aggfunc="mean").rename_axis(columns=None).index, 
              y=df5.pivot_table("PERCENTAGE_OCCUPANCY", "NEW_TIME", "KMeans Color", 
                                aggfunc="mean").rename_axis(columns=None).columns,
              color_discrete_map={
                 "blue": 'rgb(55,126,184)',
                 "green": 'rgb(77,174,74)',
                 "red": 'rgb(228,26,28)',
                 "orange":'rgb(255,127,0)',
                 "black" : '#222A2A'},
                 title="Dublin Bike Stations Cluster by Mean Occupancy % ")
fig.update_layout(
    legend_title = "Cluster",
    title_font_color="gray",
    legend_title_font_color="black",
    xaxis_title= 'Hour',
    yaxis_title= "Occupancy %"
)
# Set custom x-axis labels
fig.update_xaxes(
    ticktext=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00",
              "12:00:00","14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
    tickvals=["01:00:00", "04:00:00", "06:00:00", "08:00:00","10:00:00","12:00:00",
              "14:00:00","16:00:00","18:00:00","20:00:00","23:00:00"],
)
fig.update_xaxes(tickangle=90)
fig.show()

Lecturer Name	Module	Email address
David McQuaid	Data Preparation & Visualisation for Data Analytics	dmcquaid@cct.ie
David McQuaid	Programming for Data Analytics	dmcquaid@cct.ie
Dr. Muhammad Iqbal	Machine Learning for Data Analytics	miqbal@cct.ie
Marina Iantorno	Statistics for Data Analytics	miantorno@cct.ie

	STATION ID	TIME	LAST UPDATED	NAME	BIKE STANDS	AVAILABLE BIKE STANDS	AVAILABLE BIKES	STATUS	ADDRESS	LATITUDE	LONGITUDE
0	2	2021-04-01 00:00:03	2021-03-31 23:56:25	BLESSINGTON STREET	20	6	14	Open	Blessington Street	53.35677	-6.26814
1	2	2021-04-01 00:05:03	2021-04-01 00:00:12	BLESSINGTON STREET	20	6	14	Open	Blessington Street	53.35677	-6.26814
2	2	2021-04-01 00:10:03	2021-04-01 00:00:12	BLESSINGTON STREET	20	6	14	Open	Blessington Street	53.35677	-6.26814
3	2	2021-04-01 00:15:02	2021-04-01 00:10:17	BLESSINGTON STREET	20	6	14	Open	Blessington Street	53.35677	-6.26814
4	2	2021-04-01 00:20:03	2021-04-01 00:10:17	BLESSINGTON STREET	20	6	14	Open	Blessington Street	53.35677	-6.26814

	STATION ID	BIKE STANDS	AVAILABLE BIKE STANDS	AVAILABLE BIKES
count	2884576.0	2884576.0	2884576.0	2884576.0
mean	60.0	32.0	20.0	12.0
std	34.0	8.0	9.0	7.0
min	2.0	16.0	0.0	0.0
25%	31.0	29.0	13.0	6.0
50%	61.0	30.0	20.0	11.0
75%	90.0	40.0	27.0	16.0
max	117.0	40.0	40.0	40.0

Number of Variables	11
Number of Rows	2.8846e+06
Missing Cells	0
Missing Cells (%)	0.0%
Duplicate Rows	0
Duplicate Rows (%)	0.0%
Total Size in Memory	748.6 MB
Average Row Size in Memory	272.1 B
Variable Types	Numerical: 6 DateTime: 2 Categorical: 3

BIKE STANDS is skewed	Skewed
NAME has a high cardinality: 109 distinct values	High Cardinality
ADDRESS has a high cardinality: 109 distinct values	High Cardinality
STATUS has constant value "Open"	Constant
STATUS has constant length 4	Constant Length
LONGITUDE has 2884576 (100.0%) negatives	Negatives

	TIME	LAST UPDATED	NAME	STATUS	ADDRESS
count	2884576	2884576	2884576	2884576	2884576
unique	26464	1311282	109	1	109
top	2021-04-01 00:00:03	2021-05-26 14:33:05	BLESSINGTON STREET	Open	Blessington Street
freq	109	222	26464	2884576	26464

Approximate Distinct Count	109
Approximate Unique (%)	0.0%
Missing	0
Missing (%)	0.0%
Infinite	0
Infinite (%)	0.0%
Memory Size	44.0 MB
Mean	60.3303
Minimum	2
Maximum	117
Zeros	0
Zeros (%)	0.0%
Negatives	0
Negatives (%)	0.0%

Minimum	2
5-th Percentile	7
Q1	31
Median	61
Q3	91
95-th Percentile	112
Maximum	117
Range	115
IQR	60

Mean	60.3303
Standard Deviation	33.8648
Variance	1146.8271
Sum	1.7403e+08
Skewness	-0.03532
Kurtosis	-1.2204
Coefficient of Variation	0.5613

Distinct Count	26401.2853
Approximate Unique (%)	0.9%
Missing	0
Missing (%)	0.0%
Memory Size	22.0 MB
Minimum	2021-04-01 00:00:03
Maximum	2021-07-01 23:55:02

Distinct Count	1.3091e+06
Approximate Unique (%)	45.4%
Missing	0
Missing (%)	0.0%
Memory Size	22.0 MB
Minimum	2021-03-31 23:49:35
Maximum	2021-07-01 23:54:05

Mean	16.5596
Standard Deviation	4.7708
Median	16
Minimum	9
Maximum	33

1st row	BLESSINGTON STREET
2nd row	BLESSINGTON STREET
3rd row	BLESSINGTON STREET
4th row	BLESSINGTON STREET
5th row	BLESSINGTON STREET

Count	42792288
Lowercase Letter	0
Space Separator	4207776
Uppercase Letter	42792288
Dash Punctuation	0
Decimal Number	0

Mean	32.1101
Standard Deviation	7.6486
Variance	58.5017
Sum	9.2624e+07
Skewness	-0.4316
Kurtosis	-1.1335
Coefficient of Variation	0.2382

Minimum	53.3301
5-th Percentile	53.3337
Q1	53.3398
Median	53.3452
Q3	53.3509
95-th Percentile	53.3584
Maximum	53.36
Range	0.02988
IQR	0.01116

Minimum	-6.31
5-th Percentile	-6.2978
Q1	-6.2752
Median	-6.2632
Q3	-6.2509
95-th Percentile	-6.2372
Maximum	-6.2309
Range	0.07916
IQR	0.02433

	NAME	BIKE STANDS	AVAILABLE BIKE STANDS	AVAILABLE BIKES	DAY_NUMBER	DAY_TYPE	TIME_ROUNDED_10_MIN	NEW_TIME	PERCENTAGE_OCCUPANCY
0	BLESSINGTON STREET	20	6	14	2	Weekday	2021-04-01 00:00:00	00:00:00	0.7
1	BLESSINGTON STREET	20	6	14	3	Weekday	2021-04-01 00:00:00	00:00:00	0.7
2	BLESSINGTON STREET	20	6	14	3	Weekday	2021-04-01 00:00:00	00:00:00	0.7
3	BLESSINGTON STREET	20	6	14	3	Weekday	2021-04-01 00:10:00	00:10:00	0.7
4	BLESSINGTON STREET	20	6	14	3	Weekday	2021-04-01 00:10:00	00:10:00	0.7

Mean	20.1284
Standard Deviation	9.306
Variance	86.6023
Sum	5.8062e+07
Skewness	0.001502
Kurtosis	-0.6708
Coefficient of Variation	0.4623

Mean	11.8447
Standard Deviation	7.3748
Variance	54.3873
Sum	3.4167e+07
Skewness	0.6326
Kurtosis	0.2036
Coefficient of Variation	0.6226

Count	11538304
Lowercase Letter	8653728
Space Separator	0
Uppercase Letter	2884576
Dash Punctuation	0
Decimal Number	0

Count	42818752
Lowercase Letter	35699936
Space Separator	4207776
Uppercase Letter	7118816
Dash Punctuation	0
Decimal Number	0

Mean	53.3455
Standard Deviation	0.007569
Variance	5.7286e-05
Sum	1.5388e+08
Skewness	0.07786
Kurtosis	-0.7722
Coefficient of Variation	0.00014188

Mean	-6.2645
Standard Deviation	0.01819
Variance	0.00033076
Sum	-1.807e+07
Skewness	-0.4446
Kurtosis	-0.2628
Coefficient of Variation	-0.002903

	00:00:00	00:10:00	00:20:00	00:30:00	00:40:00	00:50:00	01:00:00	01:10:00	01:20:00	01:30:00	...	22:20:00	22:30:00	22:40:00	22:50:00	23:00:00	23:10:00	23:20:00	23:30:00	23:40:00	23:50:00
NAME
AVONDALE ROAD	0.396364	0.393462	0.390891	0.393462	0.392939	0.394922	0.393269	0.394423	0.393182	0.393893	...	0.384659	0.381481	0.379198	0.381055	0.375564	0.377037	0.377820	0.380370	0.380709	0.379960
BENSON STREET	0.233906	0.240625	0.240820	0.241538	0.241477	0.242308	0.239205	0.236538	0.241031	0.239773	...	0.239366	0.237868	0.224609	0.231298	0.227555	0.223837	0.221507	0.227539	0.223496	0.232863
BLACKHALL PLACE	0.646405	0.652116	0.657881	0.656923	0.654593	0.652344	0.655062	0.655556	0.665404	0.656250	...	0.625191	0.614734	0.626768	0.626111	0.642058	0.656331	0.661578	0.662087	0.660652	0.664428
BLESSINGTON STREET	0.451829	0.456250	0.462403	0.462214	0.458594	0.459924	0.459615	0.458209	0.460150	0.455303	...	0.454198	0.467279	0.459701	0.482377	0.467557	0.463258	0.476984	0.482197	0.466165	0.471805
BOLTON STREET	0.386538	0.381641	0.382946	0.392424	0.388976	0.387109	0.387594	0.386194	0.389474	0.385878	...	0.398047	0.396538	0.401799	0.402692	0.394203	0.412016	0.404651	0.400376	0.400373	0.410938

	First_Comp	Second_Comp	Cluster Hierarquical
0	-4.180715	-5.461058	1
1	-7.968315	1.147348	0
2	10.222895	-5.804465	0
3	-0.669812	-4.198103	0
4	-8.338002	-8.573231	1

	NAME	Cluster KMeans	LATITUDE	LONGITUDE	KMeans Color
0	AVONDALE ROAD	2	53.359406	-6.276142	green
1	BENSON STREET	4	53.344154	-6.233451	black
2	BLACKHALL PLACE	0	53.348801	-6.281637	blue
3	BLESSINGTON STREET	2	53.356770	-6.268140	green
4	BOLTON STREET	4	53.351181	-6.269859	black

MSc in Data Analytics¶

Individual Project CA1¶

Table of contents:¶

Initial Exploratory and Data Analysis¶

Overview

Dataset Statistics

Dataset Insights

Variables

STATION ID

Quantile Statistics

Descriptive Statistics

TIME

LAST UPDATED

NAME

Length

Sample

Letter

BIKE STANDS

Quantile Statistics

Descriptive Statistics

AVAILABLE BIKE STANDS

Quantile Statistics

Descriptive Statistics

AVAILABLE BIKES

Quantile Statistics

Descriptive Statistics

STATUS

Length

Sample

Letter

ADDRESS

Length

Sample

Letter

LATITUDE

Quantile Statistics

Descriptive Statistics

LONGITUDE

Quantile Statistics

Descriptive Statistics

Interactions

Correlations

Missing Values

Statistics Questions¶

Using Longitude and Latitude to visualize Dublin Bike stations in a MAP with Folium¶

Data Preparation and Feature engineering¶

Re-Shaping Data and Build M Learning Models¶

Dimensionality Reduction using Principal Components Analyses¶

Hierarchy Clustering¶

KMeans Clustering¶

Comparing models with silhouette score¶

Imputing K Means Labels and evaluating results with Data Visualization¶

Clonclusion¶

References:¶